Typical Information Stored in Index Databases

Some index databases use very special architectures, like WAIS, but most contain some of the information items listed below.

• For each document a number of words is selected, for which there is an entry in the database linking that word to the document. (This number can be as low as 20 and as high as all words from the document.)
  
  
• For each occurrence of a selected word in a document some position info is stored. (exact position, or approximate, or structural)
  
  
• For each document a small piece is stored (usually the beginning) in order to be able to present a short abstract.

Note that the above information is not sufficient to find exact matches of phrases which consist of words that are tied together. In order to find "world-wide" the database needs to store "world-wide". Storing only "world" and "wide" means that "world wide" and "world-wide" cannot be distinguished.

Most search engines offer a choice between caseless or case sensitive matching. This means that in the database upper- and lowercase must be distinguished.